159

EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

ZHE-XI LUO

Section of Vertebrate Paleontology, Carnegie Museum of Natural History,4400 Forbes Avenue, Pittsburgh, PA 15213

ZOFIA KIELAN-JAWOROWSKA

Instytut Paleobiologii PAN, ul. Twarda 51/55 PL-00-818 Warszawa, Poland

RICHARD L. CIFELLI

Oklahoma Museum of Natural History, 2401 Chautauqua,Norman, OK 73072

ABSTRACT

We provide a review of dental replacement features in stemclades of mammals and an hypothetical outline for the evolutionof replacement frequency, mode, and sequence in early mam-malian evolution. The origin of mammals is characterized by ashift from a primitive pattern of multiple, alternating replace-ments of all postcanines in most cynodonts to a derived patternof single, sequential replacement of postcanines. The stem mam-mal Sinoconodon, however, retained some primitive replacementfeatures of cynodonts. The clade of Morganucodon 1 crownmammals is characterized by the typical mammalian diphyodontreplacement in which antemolars are replaced by one generationin antero-posterior sequence, but molars are not replaced. The

stem clades of crown mammals including multituberculates andeutriconodonts have an antero-posterior sequential and diphy-odont replacement of premolars. By contrast, stem taxa of thetrechnotherian clade (Zhangheotherium, Dryolestes, and Slaugh-teria) are characterized by an alternating (p2 → p4 → p3) anddiphyodont replacement, a condition that is shared by basal eu-therians. The sequential replacements of premolars in most extantplacentals (either antero-posteriorly p2 → p3 → p4 as in ungu-lates and carnivores, or postero-anteriorly p4 → p3 → p2 as insome insectivores) would represent secondarily derived condi-tions within eutherians. The single replacement of P3/p3 of meta-therians is the most derived for all therian mammals.

INTRODUCTION

Mammals differ from their phylogenetic rela-tives—nonmammalian cynodonts—in their greatlyreduced number of successional teeth per tooth lo-cus. Living mammals have two generations of teeth(diphyodonty) at most, whereas nonmammalian cy-nodonts had multiple generations of tooth replace-ment sustained throughout life (polyphyodonty).Mammalian dental replacement is limited partly bya delay in the onset of dental eruption in neonatesthat nurse on maternal milk. During lactation, thetoothless neonates can achieve a considerableamount of cranial growth at a rapid rate before erup-tion of the first generation of deciduous teeth(Brink, 1956; Hopson, 1973; Pond, 1977; Luckett,1993). The mammalian dental replacement is alsoreduced partly by an early termination of dental re-placement related to the determinate skull growthin mammals. Prior to weaning, the rate of skullgrowth exceeds that of the postcranial skeleton. Af-ter weaning the rate of skull growth slows down.The termination of the skull growth usually coin-cides with eruption of the last molar (Pond, 1977).

Diphyodont dental replacement of mammals is amajor apomorphy because it is certainly correlatedwith the determinate growth pattern of the skull,and partly correlated with lactation, which is themost important mammalian characteristic (Pond,1977; Tyndale-Biscoe and Renfree, 1987; Jenkins,1990; and Zeller, 1999). Nursing of the neonates bymaternal milk has had a profound impact on mam-malian growth patterns and on development of nu-merous apomorphies in dentition and skull, such as:the reduced replacement of the postcanines, devel-opment of precise molar occlusion, formation of thedentary-squamosal temporomandibular joint, as firstpointed out by Brink (1956) and elaborated bymany others (Hopson and Crompton, 1969; Romer,1970; Ziegler, 1971; Hopson, 1971; Pond, 1977;Kermack and Kermack, 1984; Gow, 1985; Cromp-ton and Hylander, 1986; Luckett, 1993; Luo, 1994;Crompton, 1995).

The complex phenomenon of dental replacementcan be broken down to several basic morphologicalelements. (1) Replacement frequency—number of

160 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

successional teeth at each tooth locus, especiallywhether replacement occurred in the posterior mo-lariforms. The rate of dental replacement is directlycorrelated with the patterns of skull growth. (2)Mode of replacement: alternating versus sequential.

(3) Direction or sequence of replacement: antero-posterior versus postero-anterior. In this study wewill attempt to outline these main features of thedental replacement patterns in cynodont-mammalevolution.

DENTAL REPLACEMENT OF EXTANT MAMMALS

←

Fig. 1.—Comparison of dental replacement between nonmam-malian cynodonts and extant therian mammals. A. The nonmam-malian cynodont Pachygenelus (from Crompton and Luo, 1993:fig. 4.14). B. The eutherian Canis latrans (coyote): above—de-ciduous lower dentition of a juvenile (based on Carnegie Mu-seum CM 8653); below—permanent dentition of an adult (CM7365). C. The metatherians Alphadon (modified from Clemens,1966, and Cifelli et al., 1996). Teeth in light gray: erupting/suc-cessor teeth (5 permanent teeth after eruption). Dark gray: de-ciduous predecessor teeth yet to be replaced. Gray/hatched: de-ciduous predecessor teeth in alternate positions, yet to be re-placed. Unshaded: permanent teeth. Abbreviations: di, deciduousincisor; dc, deciduous canine; dp, deciduous premolars; dpc, de-ciduous postcanines; m, molars; pc, postcanines; rpc, replacingpostcanines. Most nonmammalian cynodonts (except gompho-donts and tritylodontids) are characterized by alternating replace-ments of all teeth for multiple generations, in correlation withindeterminate growth of the jaw. The deciduous postcanines arereplaced at every other, or every third tooth position in the samereplacement wave (as seen in Pachygenelus). As the successortooth tends to be larger than the predecessor tooth in each toothposition, the jaw had to grow in length to accommodate the larg-er replacement teeth in cynodonts, as in extant diapsids. In mostextant placental mammals, there is only one generation of re-placement for the antemolars but no replacement of molars (asseen in Canis). Tooth replacement is more reduced in metathe-rians (as seen in Alphadon), postcanine replacement only occursin the ultimate premolar P3/p3.

Placentals (Fig. 1).—The diphyodont dental re-placement in most living placental mammals ischaracterized by a single replacement of deciduous(‘‘milk’’) incisors, canines, and premolars by a sec-ond generation of permanent incisors, canines andpremolars. Molars are part of the permanent denti-tion and are never replaced (Ziegler, 1971; Williamset al., 1989; Nowak, 1991; Evans, 1995).

There are several well-known exceptions to thistypical diphyodont pattern among the placentals. Asthe antemolar replacement has some degree of ho-moplasy, it can vary among genera of the same fam-ily. In some very small soricid and talpid insecti-

vores, none of the premolars is replaced and theentire postcanine series is monophyodont (Osborn,1971; Bloch et al., 1998). The embryonic precursorsto the deciduous antemolars in the highly altricialneonates are reabsorbed before eruption (Church-field, 1990; Nowak, 1991). As a result, these smallinsectivores have only a one generation of function-al postcanine teeth (monophyodonty). The samemonophyodont condition appears to be also presentin some geolabidid insectivores of the Late Creta-ceous and early Tertiary (Lillegraven et al., 1981),although not in others (Bloch et al., 1998). Otherexceptions to the typical placental diphyodont den-tition can be seen in the anteaters (myrmecophagidxenarthrans) and the pangolins (manid pholidotans).These mammals are toothless and rely on the tongueto feed on ants and other insects. Adult mysticetewhales have baleen for filter feeding instead of teeth

2004 161LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

(Nowak, 1991). Nonetheless, most placental mam-mals have a diphyodont dentition in which all de-ciduous antemolars (including the premolars or‘‘milk molars’’) are replaced by the successionaland permanent teeth. This generalized placentalcondition is less specialized than those of extantmarsupials and monotremes.

The sequence of replacement of premolars is alsovariable among various extant placental groups. Inthe talpid insectivore Scapanus, the eruption of de-ciduous premolars and their replacement occur ear-lier in the ultimate premolar locus than the moreanterior premolar loci (Zielger, 1971). The samepostero-anterior sequences of premolar sheddingand eruption also occur in the chrysochlorid Er-emitalpa (Kindahl, 1963), the macroscelidid Ele-phantulus (Kindahl, 1957), and in some lipotyph-lans (Luckett, 1993). This postero-anterior sequenceof premolar replacement was previously consideredto be a general pattern of placental insectivores (Os-born, 1973; Luckett, 1993). Among primates, mostnon-anthropoid taxa have postero-anterior sequen-tial replacements (Smith, 2000), with the interestingexceptions of the stem anthropoid Apidium, inwhich the upper and lower premolars are replacedalternatingly (P2 → P4 → P3) (Kay and Simons,1983). Australopithecines shows an alternating pat-tern of upper premolar replacement (Smith, 1994).Humans at 10 years of age show also an alternatingpattern as P2 breaches the gum before P1 and P3(Williams et al., 1989:fig. 8.84). Primates as a wholehave postero-anterior sequential premolar replace-ments (Smith, 2000).

However, the sequence of premolar replacementsin placental ungulates, carnivores, and other placen-tal stem taxa are in the opposite, antero-posteriordirection (Schmid, 1972; Smith, 2000). The decid-uous premolars are replaced from the anterior to themore posterior positions in horse, sheep, pig, anddomestic dog. In cattle (Bos), dp2 and dp3 are re-placed about the same time (Olsen, pers. comm.)and earlier than dp4. The erupting sequence of p3erupting before p4 (in an antero-posterior direction)for the premolars is known for the paromomyid Aci-domomys (Bloch et al., 2002), a stem placental tax-on. In the Tertiary stem placental Microsyops, p4erupts before p3 and in Plesiadapis p3 and p4 eruptabout the same time (Bloch et al., 2002). But the

latter two taxa do not seem to represent the gener-alized condition of most early Tertiary mammals.Given the distribution of these characteristics, it isclear that primitively the placental crown group asa whole has sequential replacement of premolars,but it is not clear whether the antero-posterior di-rection, or the postero-anterior direction would pre-sent the ancestral condition of all extant placentals.

Metatherians (Fig. 1C).—In metatherians includ-ing extant marsupials, only the ultimate premolar(P3/p3) has both an erupted deciduous tooth and asuccessor tooth (Luckett, 1993; Cifelli et al., 1996;Rougier et al., 1998). Other antemolars have onlyone generation of functional teeth in most marsu-pials. The first and second premolars have the erupt-ed deciduous teeth retained in adults, without re-placing successors. The incisors and canines haverudimentary and nonerupting deciduous predeces-sors that are lost or reabsorbed, followed by theaccelerated development of successional permanentteeth (Luckett, 1993; Luckett and Woolly, 1996;Luckett and Hong, 2000).

The greatly reduced diphyodont replacement ofmarsupials is correlated with their perinatal special-ization for prolonged lactation by fixation on thematernal nipples (Luckett, 1977, 1993; Tyndale-Biscoe and Renfree, 1987; Maier, 1993; Zeller,1999). This pattern is now traced back to some ofthe earliest-known stem taxa of metatherians of theLate Cretaceous, by the studies of Clemens (1966),Lillegraven (1969), Cifelli et al. (1996), and Roug-ier et al. (1998) that shed light on the timing of theevolution of the marsupial developmental speciali-zations.

Monotremes.—The echnidas (Tachyglossidae)have no teeth and rely on the tongue for feeding.The platypus Ornithorhynchus (Ornithorhynchidae)has deciduous teeth in early growth stages, but theseare replaced by horny dental pads in adults (Simp-son, 1929; Green, 1937; Woodburne and Tedford,1975). Of the two deciduous premolars, dp1 has noreplacing successor whereas dp2 has a successionaldental lamina (Luckett and Zeller, 1989). In this re-gard, the premolar replacement in Ornithorhynchusretains a vestige of the diphyodont replacement pre-molar of therian mammals (Green, 1937; Parring-ton, 1974; Luckett and Zeller, 1989; Archer et al.,1993).

162 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

Fig. 2.—Continual dental replacement during cranial and mandibular growth in the stem mammal Sinoconodon (from Zhang et al.,1998 and Luo et al., 2001). A. Upper dental replacement. B. Lower dental replacement and growth of mandible. C. Ranges of skullsize of currently known specimens in Sinoconodon and Morganucodon, and estimate of their body mass on the basis of extant insectivorescaling data. At least four generations of replacement canines occurred throughout the entire observed size range of Sinoconodon skullscollected from the local Early Jurassic Lower Lufeng Formation. The postcanine replacement shows an antero-posterior sequence: P1→ P2 → ‘‘M3’’ → ‘‘M4’’ → ‘‘M5’’. In correlation with the continuous replacement of smaller predecessor postcanines by largersuccessors, the jaws and skulls increased in size (the skull length ranging from 22 to 62 mm; estimated body masses ranged from ;13gto ;517g) in Sinoconodon. By comparison, Morganucodon skulls from the same Lufeng fauna have ranged from 27 to 38 mm (withestimated body masses ranging from 27g to 89g). This suggests that Sinoconodon had continuous skull growth accompanied by con-tinuous tooth replacement well into late stage of adult life, a characteristic of indeterminate growth pattern of extant nonmammalianamniotes and nonmammalian cynodonts. The narrower range of growth in Morganucodon suggests that it is closer to determinate growthof extant mammals than Sinoconodon and nonmammalian cynodonts. Solid triangles: Morganucodon and Sinoconodon skulls; opencircles, extant insectivore skull and body mass data in regression of Y53.68X-3.83 (data from Gingerich and Smith, 1984).

2004 163LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

NONMAMMALIAN AMNIOTES

Diapsids.—Most of the toothed diapsid amnioteshave multiple generations of teeth, or polyphyodontdental replacements (Edmund, 1960; Hildebrand,1974). In extant diapsids, such as lizards, hatchlingsmust be capable of independent feeding with a fullset of functional teeth. As demonstrated by Edmund(1960), Osborn (1971, 1973; 1974a), and Wester-gaard and Ferguson (1986, 1987) the teeth eruptearly in growth and are subsequently replaced con-tinuously throughout life. The tooth replacement oc-curs alternately at every other, or every third toothposition (Edmund, 1960; Osborn, 1971). This con-tinual and alternating dental replacement pattern ap-pears to be primitive for extant tetrapods and am-niotes as a whole (Hildebrand, 1974; Berkowitz,2000). Continual tooth replacement is correlatedwith indeterminate growth of the jaws, because thecontinued lengthening of the jaws is required forthe smaller predecessor teeth to be replaced by thelarger successor teeth that are fixed in size and mor-phology by enamel on the crown at eruption. Thecontinual replacement by larger teeth in older in-dividuals with simultaneous lengthening of the jawsis characteristic of the indeterminate growth inmany diapsids (Osborn, 1971; Westergaard and Fer-guson, 1986, 1987) including dinosaurs (Erickson,1996; Erickson and Tumanova, 2000).

Nonmammalian synapsids (Fig. 1A).—Most cy-nodonts are characterized by alternate, multiple den-tal replacements and the related indeterminate jawgrowth, as are most extant diapsids. Alternating andcontinuous dental replacements have been docu-mented for Thrinaxodon (Parrington, 1936; Cromp-

ton, 1963; Osborn and Crompton, 1973) and Pach-ygenelus (Gow, 1980; Crompton and Luo, 1993),among other nonmammalian cynodonts. Postca-nines in Thrinaxodon are replaced up to three timesat the posterior loci (Osborn and Crompton, 1973).Although lower than those of living diapsids, thenumber of successional teeth per tooth locus ismuch higher than for any mammals.

Two well-known exceptions to the general cy-nodont replacement pattern are diademodontids andtritylodontids. In diademodontids with the ‘‘gom-phodont’’ postcanines, and possibly also in relatedtraversodontids, the postcanine replacement is se-quential in an antero-posterior direction (Crompton,1963, 1972; Hopson, 1971; Osborn, 1974b). In tri-tylodontids, there is no replacement of postcanines.The postcanines erupt by sequential addition at theposterior end of tooth row, and the worn postca-nines are shed anteriorly (Kuhne, 1956; Hopson,1965; Cui and Sun, 1987). Nonetheless, the alter-nating and multiple replacements of Thrinaxodonare representative conditions for nonmammalian cy-nodonts as a whole.

The majority of cynodonts had the reptilian in-determinate growth because their jaws continued tolengthen as smaller predecessor teeth were replacedby larger successor teeth at the posterior loci inlarge individuals (Osborn and Crompton, 1973).Because dental replacement begins in the smallestknown individuals in cynodonts (Crompton, 1963;Hopson, 1971; Osborn, 1973; Osborn and Cromp-ton, 1973; Crompton and Luo, 1993), it is unlikelythat nonmammalian cynodonts developed lactation.

STEM CLADES OF MAMMALS

Sinoconodon (Fig. 2).—Sinoconodon showed dif-ferentiation of premolars from molars and a singlereplacement of premolars (Zhang et al., 1998; con-tra Crompton and Luo, 1993). Both are derived fea-tures of modern mammals. Replacement of the post-canines in Sinoconodon followed an antero-post-erior sequence. This is a derived character of mam-mals (defined as common ancestor of Sinoconodonand the crown Mammalia; 5 Mammaliaformes sen-su Rowe, 1988; McKenna and Bell, 1997), as itdiffers from the alternating replacement in mostnonmammalian cynodonts, although it has some re-semblance to the replacement of posterior postca-nines in diademodontids.

However, other dental replacement characteristicsof Sinoconodon are very primitive for mammals.Incisors in Sinoconodon show alternate replacementpattern. The canines were replaced at least threetimes (Crompton and Luo, 1993; Zhang et al.,1998), similar to the multiple alternating replace-ments of incisors and canines of many nonmam-malian cynodonts.

The youngest known specimens of Sinoconodon(Fig. 2) have two deciduous premolars. These de-ciduous premolars were replaced once, but theirsuccessor teeth are subsequently lost in the larger(thus older) specimens of their growth series. Theanterior molars (M1–M2) of the smaller individuals

164 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

were probably lost without replacement. The loss ofanterior postcanines (premolars and molars) has re-sulted in a postcanine diastema that is increasinglylarger in older individuals. The loss of anterior post-canines, coupled with successive addition of newlyerupted molariforms at the posterior end of the toothrow, has resulted in the posterior shift of the func-tional tooth rows in the jaws. These are primitivefeatures shared by Thrinaxodon, Probainognathus,diademodontids, and tritylodontids.

This pattern of losing the anterior postcanines ispresent, although to a lesser extent, in Morganu-codon, Hadrocodium, and possibly also in Kueh-neotherium (Kermack et al., 1968; Parrington,1971; Mills, 1971, 1984; Luo et al., 2001a). Fewother mammals have this pattern of losing anteriorpostcanines during growth, except where it presum-ably was acquired secondarily, as in proboscideansof the Tertiary.

The posterior molars (M3–M5) have one replace-ment in the larger (presumably older) individuals ofSinoconodon (Fig. 2). The replacement of a smallerultimate molariform by a larger successor followedby eruption of yet another ultimate molar in an old-er individual occurs only in Sinoconodon and dia-demodontids (Crompton, 1963; Hopson, 1971). Sin-oconodon also resembles diademodontids in thatsmaller deciduous postcanines with simpler crownsare replaced by erupting postcanines that are largerand more molariform. By contrast, in the cynodontThrinaxodon, the predecessor postcanine tends to bemore molariform and complex and the successivereplacing teeth are progressively simpler (Osbornand Crompton, 1973). A convergent pattern canalso be found in most mammals (except Gobicon-odon), in which a deciduous molariform predeces-sor is usually replaced by a permanent premolarwith less molariform crown.

Sinoconodon lacks precise dental occlusion be-cause it does not have the one-to-one oppositionbetween upper and lower molars (Crompton andSun, 1985). This is correlated with the partial re-placement of posterior molariforms and with thesuccessive posterior shift of the functional postca-nine row as a part of the indeterminate growth pat-tern of the skull (Zhang et al., 1998).

The currently available specimens of Sinocono-don show a large range of growth from the smallestindividual with an estimated body mass of about 13grams to the largest individual with an estimatedbody mass of more than 500 grams (Fig. 2C). Dur-ing this growth, the posterior molariforms were be-ing replaced while the upper and lower jaws con-tinued to lengthen in the older individuals.

From these characteristics of Sinoconodon we in-fer that it experienced indeterminate growth in itsskull, accompanied by continual tooth replacement,as in nonmammalian cynodonts (Crompton, 1963;Hopson, 1971; Osborn and Crompton, 1973; Os-born, 1974b) and modern diapsid reptiles (Osborn,1971, 1974a; Westergaard and Fergusen, 1986,1987; Berkowitz, 2000). Also it is plausible thatSinoconodon did not develop lactation because thepolyphyodont replacement of the anterior teeth hadalready begun with the smallest known individual(Fig. 2A, B) (Zhang et al., 1998).

Sinoconodon is considered to be the sister taxonto all other mammals (Crompton and Sun, 1985;Crompton and Luo, 1993; Rowe, 1993; Luo, 1994;Hopson, 1994; Wible et al., 1995; Rougier et al.,1996; Luo et al., 2001a, b). Given its position inthe cynodont-mammal phylogeny, and given itsmany primitive replacement features of nonmam-malian cynodonts, it is parsimonious to regard thedental replacement of Sinoconodon as an interme-diate stage in the character evolution from the prim-itive pattern of polyphyodont replacement seen inmost cynodonts to the derived diphyodont replace-ment of mammals. The coexistence of the mam-malian dentary/squamosal joint and the cynodont-like multiple replacements of the incisors and ca-nines in Sinoconodon indicates that the mammaliantemporomandibular joint evolved before the typicalmammalian dental replacement pattern (Luo, 1994).The replacement features in Sinoconodon also sug-gest that the reduction in the postcanine replacementpreceded the reduction of the replacement of inci-sors and canines. The suppression of dental replace-ment had occurred in the postcanines before it didin the anterior dentition.

Morganucodon and Haldanodon.—Morganuco-don had a single replacement of the posterior pre-molars (Mills, 1971; Parrington, 1971, 1973, 1978;Clemens and Lillegraven, 1986). Parrington (1971,1973, 1978) suggested that Morganucodon (referredto as ‘‘Eozostrodon’’) had a typically mammalianreplacement of the incisors and canines. Subsequentobservations are consistent with this suggestion(Kermack et al., 1973; 1981; Gow, 1985; Cromptonand Luo, 1993). The mode of replacement of pre-molars is sequential in the antero-posterior direc-tion, similar to that of Sinoconodon.

It is more difficult to document the nonreplace-ment of molariforms in morganucodonts. Gow(1986) suggested that Megazostrodon, which isclosely related to Morganucodon, might have re-placed its m2 based on the substantial evidence that

2004 165LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

this tooth is less worn than the adjacent m1 and m3.Because Megazostrodon is represented only by twospecimens (Crompton, 1974; Gow, 1986), its samplesize is too small to be certain about this. Parrington(1971) described the groove for the replacing dentallamina in some mandibles of Morganucodon. Thisgroove is usually present in cynodonts with ongoingpostcanine replacement. However, dissecting theposterior mandible did not reveal any replacementteeth, as shown by Parrington (1971, 1973). Theultimate molars in Morganucodon are variable insize, which was considered by Parrington (1973) tobe due to dimorphic or polymorphic variation.

The consensus of those who studied the dentitionof Morganucodon watsoni is that the molars werenot replaced (Mills, 1971; Crompton, 1972; Par-rington, 1973; Crompton and Parker, 1978; Clemensand Lillegraven, 1986; Luo, 1994). There is littleevidence for replacement of ultimate molars fromthe series of complete mandibles of Morganucodonoehleri (Young, 1982; Crompton and Luo, 1993)and Dinnetherium (Jenkins et al., 1983; Cromptonand Luo, 1993). The X-ray of two Morganucodonoehleri skulls shows no replacement successors tothe ultimate and penultimate functional molars. It issafe to suppose that Morganucodon achieved thetypical mammalian diphyodont dental replacementboth in antemolars and the anterior molars. It isbeyond question that Morganucodon has developedthe typical diphyodont pattern in all of its antemo-lars and its anterior molars. It is reasonable to hy-pothesize that ultimate molars in morganucodontanswere not replaced, until this hypothesis is falsifiedby additional data from a more complete samplingof skull growth series. If the variation in size andmorphology of posterior molars of Morganucodonis shown to be correlated with growth stages, thenthe mode of the skull growth of Morganucodonmust be reconsidered. With these caveats, we ten-tatively accept that the ultimate molar was not re-placed. This working hypothesis of nonreplacementof ultimate molars (Parrington, 1971; Kermack andKermack, 1984; Crompton and Luo, 1993; Luo,1994) is consistent with the skull growth pattern ofMorganucodon from the currently available (al-though still small) sample of specimens (Fig. 2C).

Morganucodon shows far smaller size range of theskulls than the contemporary Sinoconodon (Fig. 2C).The eight relatively complete skulls of Morganuco-don discovered so far range in length from 27 to 38mm (Young, 1982; Crompton and Luo, 1993; Luoet al., 1995; Zhang et al., 1998). This corresponds toa body-size range from 27 grams to 89 grams (Luoet al., 2001b), in strong contrast to a much widerrange of variation in the estimated body size of Sin-oconodon from 13 grams to over 500 grams. Thisindicates that the adult skulls grew far less in Mor-ganucodon than in Sinoconodon. Most maxillariesand mandibles of Morganucodon from the Rhaeto-Liassic fissure deposits of the Great Britain areadults, and few are juveniles (Kermack et al., 1973,1981; Parrington, 1971, 1978). Based on this Gow(1985) suggested that Morganucodon probably hada very short juvenile stage, a view supported by Luo(1994). The skull growth series and dental replace-ment indicate that Morganucodon had a short growthperiod to reach its adult size, at least by comparisonto Sinoconodon (Fig. 2C) and cynodonts. The growthpattern of Morganucodon is closer to the determinategrowth pattern typical of extant mammals than thoseof Sinoconodon and nonmammalian cynodonts. Wehereby hypothesize that Morganucodon achieved thedeterminate growth of the skull characteristics of liv-ing mammals.

Dental replacement of the docodonts was studiedby various authors since the 1920s (Simpson, 1928;Butler, 1939), but only recently could a definitivereplacement pattern be documented in an unambig-uous manner by a large sample of juvenile and adultjaws of Haldanodon (Krusat, 1980; Lillegraven andKrusat, 1991; Martin and Nowotny, 2000; Nowotnyet al., 2001). Molars were not replaced in Haldan-odon. The incisors, canines and premolars in Hal-danodon show diphyodont replacement that pro-ceeded from the front to the back in a sequentialmode (Martin and Nowotny, 2000; Nowotny et al.,2001). This is an apomorphy in comparison to Sin-oconodon and cynodonts. Haldanodon is more de-rived than the cynodont Thrinaxodon, Sinoconodon,Morganucodon, Hadrocodium, and Kuehneother-ium in lacking the increasingly large postcanine di-astema in the larger and older individuals due to theloss of the anterior premolars.

STEM CLADES OF MAMMALIAN CROWN GROUP

Gobiconodontids (Fig. 3A).—Gobiconodontidsreplaced their anterior molariform postcanines (Jen-kins and Schaff, 1988; Wang et al., 2001). Gobi-

conodon is unique in that the successor permanenttooth is similar to its deciduous predecessor in thecomplexity of molariform morphology. In most oth-

166 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

Fig. 3.—Sequential replacement of the anterior postcanines in multituberculates and gobiconodontids. A. Growth series of Gobiconodon(adopted from Jenkins and Schaff, 1988). At least two anterior molariform postcanines were replaced each by a molariform successor.The replacement shows an antero-posterior sequence: p4 (only shedding) → ‘‘m1’’ → ‘‘m2’’ → ‘‘m3’’. B. Multituberculate dentalreplacement shows an antero-posterior sequence (adopted from Greenwald, 1988): i1 → i2 → i3 → p1 → p2 → p3 → p4. Gobicono-dontids and the majority of multituberculates and retain the primitive antero-posterior sequence for replacement of premolars. Teeth inlight gray: erupting/successor teeth (5 permanent teeth after eruption). Dark gray: deciduous predecessor teeth yet to be replaced.Unshaded: permanent teeth.

er mammals, the replacing and successional tooth issimpler in crown morphology than its deciduouspredecessor at the same premolar locus. By widelyaccepted convention for defining molars versus pre-molars (e.g., Romer, 1970; Hildebrand, 1974; Wil-liams et al., 1989; Evans, 1995), the postcanine lociwith replacement should be considered to be pre-molars, regardless whether the permanent tooth ofa premolar locus has a molariform crown, or a pre-molariform crown. The wave of replacement of theanterior postcanines is sequential, continuous fromthe front to the back, as shown by Jenkins and

Schaff (1988), and corroborated by Wang et al.(2001). Based on limited evidence, Simpson (1928)interpreted that, in Triconodon, p1–p3 had been re-placed before the replacement of dp4 by p4. If thiscan be corroborated, then Triconodon also has an-tero-posterior sequence of replacement of the pre-molars, as in gobiconodontids. However, Tricono-don has a much simpler permanent p4 than its fullymolariform dp4, and this differs from gobiconodon-tids in which both the permanent (successor) andthe deciduous (predecessor) teeth at the same locusare equally molariform.

2004 167LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

Multituberculates (Fig. 3B).—The incisors and atleast some premolars are diphyodont in their re-placement (Szalay, 1965; Hahn, 1978; Clemens,1963; Clemens and Kielan-Jaworowska, 1979;Greenwald, 1988; Hahn and Hahn, 1998). Green-wald (1988) proposed that multituberculates of theNorth American Tertiary, especially Taeniolabis,have a diphyodont replacement similar to that seenin most placental mammals. In these multitubercu-lates, tooth eruption and replacement occurred in anantero-posterior sequence. Hahn and Hahn (1998)believed that the Late Jurassic paulchoffatiid Kie-lanodon is one exception to the typical multituber-culate pattern in which the replacement of the pre-molars could have occurred in an alternating mode,in two waves, and in the postero-anterior direction,as in Thrinaxodon.

The differences between the pauchoffatiid Kie-lanodon (Hahn and Hahn, 1998) and Tertiary mul-tituberculates (Greenwald, 1988) can be interpretedby two scenarios. First, the postero-anterior and al-ternate replacement in paulchoffatiids may be thebasal condition of all multituberculates, as proposedby Hahn and Hahn (1998). If so, then multituber-culates ancestrally would bear some resemblance toThrinaxodon and to trechnotherians (more discus-sion below). The antero-posterior and sequential re-placement observed in the North American Tertiarymultituberculates would therefore represent a sec-ondarily derived condition.

The second possible interpretation is that the an-tero-posterior sequential replacement is primitivefor multituberculates as a group as proposed byGreenwald (1988), and the alternating replacementof paulchoffatiids is not only atavistic to the distantcynodonts, but also convergent to that of trechnoth-erians (to be described below). According to Green-wald’s (1988) interpretation, multituberculateswould be more similar to Sinoconodon, Morganu-codon, Haldanodon and eutriconodonts, than totrechnotherians, in the replacement sequence.Which of these conditions is basal to multituber-culates as a whole depends on the position of mul-tituberculates on the mammalian phylogenetic tree(see the recent reviews on multituberculate relation-ships by Butler, 2000 and Luo et al., 2002). Green-wald’s (1988) interpretation is consistent with amore parsimonious explanation of evolution of den-tal replacement among major clades of mammals.Here we tentatively accept that antero-posterior se-quential replacement of antemolars is characteristicof multituberculates as a whole (sensu Greenwald,1988).

Trechnotherians (sensu McKenna, 1975; McKen-na and Bell, 1997).—The majority of stem therianshas an alternating pattern in an antero-posterior se-quence for their diphyodont premolar replacement(Fig. 4). The alternating replacement in trechnoth-erians is limited to only one generation of successorper tooth locus. By contrast, the alternating replace-ment in many cynodonts is in the postero-anteriordirection and with multiple generations of succes-sors.

The spalacotheriid ‘‘symmetrodont’’ Zhangheo-therium has an alternate diphyodont replacement inan antero-posterior sequence in the lower jaws (Fig.4A). The currently available juvenile and subadultspecimens show that permanent p1 erupted first, fol-lowed then by shedding of dp3 and eruption of per-manent p3, and lastly by replacement at the p2 lo-cus. The replacement at the p3 locus occurs aroundthe time of eruption of m5. The replacement at thep2 locus occurs around the time of eruption of m6.Thus the sequence of replacement is: p1 → p3 →p2, both for the shedding of deciduous teeth and foreruption of permanent teeth. We suggest that thisalternate premolar replacement in Zhangheotheriumcould be applicable to other spalacotheriids forwhich replacement of premolars is known (Cifelli,1999). In all premolar loci, the deciduous predeces-sor tooth is more ‘‘molariform’’ than its permanentsuccessor. There is no evidence that molars are re-placed in any spalacotheriids.

Dental replacement in the ‘‘eupantothere’’ Dry-olestes is documented with extensive data (Martin,1997, 1999). Deciduous teeth have long been rec-ognized in dryolestids (Butler, 1939; Butler andKrebs, 1973). Martin (1997, 1999) further demon-strated that all antemolar teeth are replaced in Dry-olestes. The diphyodont replacement, at least in thelower jaw, occurs in the antero-posterior sequenceby two waves. The first replacing wave consists ofi2, i4, p1, p3, which are followed by the secondwave of i1, i3, c, p2, p4. p4 is the last premolar toerupt, just prior to eruption of the sixth molar (m6).The premolar replacement (both shedding of decid-uous teeth and eruption of permanent teeth) is char-acterized as: p1 → p3 → p2 → p4 (Martin, 1997).

The stem boreosphenidan Slaughteria (referred toas a ‘‘tribothere’’ by many previous authors) alsohas a similar replacement sequence of premolars:p3 → p2 → p4 (Kobayashi et al., 2002). Therefore,the alternating sequence of premolar replacement(p3 → p2 → p4) is consistently present from spa-lacotheriids through dryolestids to stem northern tri-

168 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

Fig. 4.—Alternating premolar replacement in stem therians and some eutherian mammals. A. The ‘‘symmetrodont’’ (stem trechnotherian)Zhangheotherium (subadult adopted from Hu et al., 1997, 1998; juvenile is based on National Geological Museum of China, NGMC352, from Luo et al., 2001c). B. The ‘‘eupantothere’’ (stem cladotherian) Dryolestes (adopted from Martin, 1999). C. The stem bor-eosphenidan (a ‘‘tribothere’’) Slaughteria (adopted from Kobayashi et al., 2002). D. The eutherian Kennalestes (modified from Kielan-Jaworowska, 1981). E. The eutherian Daulestes (from McKenna et al., 2000). F. The metatherian Alphadon (modified from Clemens,1966, and Cifelli et al., 1996). Teeth in light gray: erupting/successor teeth (5 permanent teeth after eruption). Dark gray: deciduouspredecessor teeth yet to be replaced. Gray/hatched: deciduous predecessor teeth in alternate positions, yet to be replaced. Unshaded:permanent teeth. By comparison to the antero-posterior sequential replacement of premolars gobiconodontids, multituberculates andSinoconodon, the premolar replacement in trechnotherians (clade of Zhangheotherium and extant therians) occurred alternately, char-acterized by P1 → P3 → P2 → P4. This pattern was retained in basal eutherians (Kennalestes and Daulestes). It is only in the placentalcrown group (such as in carnivores, ungulates and primates) that the premolar replacement occurs sequentially, which must be consideredto be a secondarily derived condition.

bosphenic mammals in three different hierarchies oftherian mammal phylogeny, making a compellingcase that this is a widespread condition for all basaltrechnotherian lineages (including the basal euthe-rians) for which the replacements have been known(Fig. 5). This is different from the primitive patternof the antero-posterior sequential replacement of

postcanines in all mammal groups outside the trech-notherian clade, such as multituberculates (exceptKielanodon), Gobiconodon, Haldanodon, Morgan-ucodon, Sinoconodon, and possibly triconodontids.We regard the alternating replacement of premolars(at least for p3 → p2 → p4) to be a synapomorphyof the trechnotherian clade.

2004 169LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

Fig. 5.—Acquisition of apomorphies in dental replacement and skull growth pattern in early evolution of mammals. The plesiomorphiccondition of primitive nonmammalian cynodonts: multiple (polyphyodont) and alternate replacements of all teeth. Acquisition of apo-morphies: Node 1 (mammals 5 mammaliaforms of Rowe, 1988; McKenna and Bell, 1997): Differentiation of premolars and molars, asingle replacement of premolars, a single replacement of some posterior molariforms in antero-posterior sequence. Node 2 (clade ofMorganucodon and crown mammals): single replacement of incisors and canines, antero-posterior sequence for a single replacement ofpremolars, molars without replacement; determinate skull growth pattern; the diphyodont replacement of incisors and canines is possiblyassociated with lactation. Node 3 (trechnotherian): alternating single replacement of premolars, a pattern that is shared by the knownstem eutherians. Node 4 (extant placentals): reversal to the antero-posterior sequential replacement of premolars. Node 5 (metatherians):nonreplacement of most teeth (except P3/p3), which represents the most derived feature of dental replacement among mammals.

STEM CLADES OF EUTHERIANS AND METATHERIANS

Stem Eutherians (Fig. 4D, E).—The most prim-itive eutherians known for their full dentition, Pro-kennalestes and Eomaia, have five premolars (Kie-lan-Jaworowska and Dashzeveg, 1989; Sigogneau-Russell et al., 1992; Ji et al., 2002). The more de-rived zhelestids also have five premolars (Archibaldand Averianov, 1997; Nessov et al., 1998; Cifelli,2000). But the replacement pattern is unknown inthese taxa. The more derived eutherians have fourpremolars, and it is generally accepted that the ho-molog of the third premolar (P3/p3) of Prokenna-lestes was lost in the Late Cretaceous eutherianswith four premolars (McKenna, 1975; Novacek,1986; Cifelli, 2000).

Replacement of anterior dentition is known forthree Late Cretaceous eutherians that have four per-manent premolars: Daulestes (McKenna et al.,2000), Kennalestes (Kielan-Jaworowska, 1975,1981) and Gypsonictops (Lillegraven, 1969; Clem-ens, 1973; Novacek, 1986). Among these, Daulestes

and Kennalestes have preserved evidence for an al-ternate replacement of premolars in an antero-pos-terior sequence as in the stem ‘‘therians’’ Zhan-gheotherium, Dryolestes and Slaughteria.

The juvenile specimen of Daulestes has four pre-molar loci, permanent p2 has erupted before per-manent p3, which in turn erupts before p4 (Fig. 4E).The sequence of eruption of the permanent teeth inDaulestes is consistent with an antero-posterior se-quential replacement. However, if taking into ac-count that permanent p2 and deciduous dp2 co-existwith each other because permanent p2 failed to dis-lodge dp2 (also see discussion by Luckett, 1993 onKennalestes), then there would be a different char-acterization of the replacement. The shedding ofdp3 occurs before the shedding of dp2, forming ashedding sequence of dp3 → dp2 → dp4.

McKenna (1975) considered that the Cretaceouseutherian Kennalestes had five premolars in juve-niles and four in adults (also see Bown and Kraus,

170 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

1979). But Luckett (1993) reinterpreted that one ofthe five premolars in the Kennalestes juvenile is aretained dp2, and that Kennalestes should have onlyfour premolars in both juvenile and adult. This in-terpretation was accepted by McKenna et al. (2000)and by Cifelli (2000), and it will be followed here(Fig. 4D).

The only known juvenile specimen of Kennales-tes also has the dP3 → dP2 → dP4 shedding se-quence of the deciduous teeth (Fig. 4D), as revealedby the comparison of the juvenile skull to the adultskull (Kielan-Jaworowska, 1975; Kielan-Jaworows-ka, 1981; Luckett, 1993). The permanent P3 eruptedlong after the permanent P2 and P4 had becomefully functioning (Fig. 4). So the erupting sequenceof these permanent molars is likely to be: P2 → P4→ P3. Luckett (1993) interpreted that the premolarbefore the erupting successional P3 was a deciduousdp2 in the Kennalestes juvenile specimen (Fig. 4),and this was accepted by McKenna et al. (2000)and Cifelli (2000). Following Luckett’s interpreta-tion, the sequence for shedding deciduous teeth inKennalestes was dp3 → dp2 → dp4. Therefore thereplacement sequence, both for shedding deciduousteeth and for erupting permanent teeth, was identi-cal to the alternating antero-posterior replacementin Slaughteria (Kobayashi et al., 2002), Dryolestes(Martin, 1997), and Zhangheotherium (Luo et al.,2001c).

The alternating, antero-posterior eruption of per-manent teeth in Kennalestes, as well as the antero-posterior alternate shedding of deciduous teeth forboth Daulestes and Kennalestes are different from

the pattern in extant placentals. In extant placentalinsectivores, the development of the precursor den-tal lamina (Luckett, 1993) and the eruption of thedeciduous premolars occur first in the ultimate pre-molar and then proceed anteriorly in talpids (Zie-gler, 1971; Osborn, 1971, 1973; Osborn and Cromp-ton, 1973), chrysochlorids (Kendahl, 1963; Osborn,1973), macroscelidids (Kadahl, 1957), some lipo-typhlans (Luckett, 1993) and fossil non-anthropoidprimates (Smith, 2000). Most eutherian carnivoresand ungulates have an antero-posterior sequence foreruption of permanent premolars (Schmid, 1972;Smith, 2000). However, from the currently availablefossil evidence of the stem eutherians, this sequen-tial premolar replacement in extant placentals is sec-ondarily derived. The alternating, antero-posteriorreplacement of antemolars of Cretaceous eutheriansis the primitive condition for eutherians as a whole.

Metatherians (Fig. 1C).—The perinatal adapta-tion in marsupial neonates by fixation of the mouthto the maternal nipple for prolonged lactation (Tyn-dale-Biscoe and Renfree, 1987; Maier, 1993; Zellerand Freyer, 2001) is directly correlated to the ac-celerated growth of the skeletomuscular system ofthe skull (Maier, 1993) relative to the brain (Smith,1996, 1997). This is also correlated with the highlytransformed diphyodont replacement of marsupialsthat is limited to a single replacement of the ulti-mate premolar P3/p3 (Luckett, 1977; 1993). Dentalreplacement in marsupials is the most derived of allmammalian tooth replacement patterns. It can betraced to the stem taxa of metatherians of the LateCretaceous (Clemens, 1966; Lillegraven, 1969; Ci-felli et al., 1996; Rougier et al., 1998).

EVOLUTION OF DENTAL REPLACEMENT FROM CYNODONTS TO MAMMALS

Diphyodont dental replacement among extanttherian mammals is characterized by several fea-tures, including two generations of ante-molars anddeciduous predecessor being more complex (‘‘mo-lariform’’) than the permanent successor at the samepremolar locus, and replacement of premolars pro-ceeding either in postero-anterior direction (smallplacental insectivores), or in antero-posterior direc-tion (placental ungulates and carnivores). It hasbeen widely accepted that the diphyodont replace-ment in therians probably was derived from the an-cestrally polyphyodont replacement of cynodontsby slowing down of the tooth replacements (see thehistorical reviews by Kermack, 1963, 1967; Osborn,1971, 1973; Hopson, 1971; Crompton and Parker,1978; Kermack and Kermack, 1984; Berkowitz,

2000). However, before the late 1980s, the detailedinformation on the number of successional teeth perpremolar loci, mode, and direction of premolar re-placement waves was not available for a wide di-versity of Mesozoic mammals, thus limiting ourprevious understanding about the evolutionary pat-terns of these replacement features.

Since the late 1980s, especially since the startlingdiscovery of the molariform replacements in Gob-iconodon (Jenkins and Schaff, 1988), a series ofnew fossils have revealed some crucial pieces ofinformation in all morphological elements of dentalreplacements in various stem mammal groups in-cluding the fossil taxa of crown mammals: (1) thereplacement frequency: number of replacements pertooth locus; (2) the replacement mode: alternating

2004 171LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

versus sequential; and (3) the direction: antero-pos-terior versus postero-anterior replacement sequence.The new data, as summarized here, have shown amuch greater degree of homoplasy in the modes andsequences of dental replacement than in the changeof replacement frequency.

Frequency of replacement (number of succesion-al teeth per locus) through the cynodont-mammaltransition shows a two-step reduction, firstly in thepremolars of Sinoconodon (Fig. 5:node 1), then inthe incisors, canines and posterior molariforms (5molars) of Morganucodon and Haldanodon (Fig. 5:node 2). The incisors and canines are also limitedto one replacement as in most extant therians, in-stead of multiple replacements as in Sinoconodon(Crompton and Luo, 1993; Zhang et al., 1998) andcynodonts. Once the apomorphic diphyodont re-placement and determinate skull growth occurred inthe stem mammals of the Early Jurassic, there havebeen no documented cases of derived mammaliangroup reversing to the polyphyodont replacementsand the indeterminate skull growth. This is the mostfundamental transformation in mammalian biologi-cal adaptation as this significant shift in skullgrowth pattern is probably correlated with the originof lactation (Brink, 1956; Hopson, 1973; Pond,1977; Gow, 1985; Luo, 1994; Zhang et al., 1998).

The mode and direction of the replacement waveshave some degree of homoplasy in cynodont-mam-mal evolution. The alternating mode of replacementis a general pattern for many cynodonts includingtritheledontids (again, except for diademodontidsand tritylodontids). If tritheledontids are the sister-taxon to mammals, as is preferred by the majorityof workers, then the mammalian origin can be char-acterized by a shift from primitive multiple, antero-posterior and alternating replacements of all post-canines in most cynodonts to a derived pattern ofsingle, and antero-posterior sequential replacementof postcanines in the clade of Sinoconodon and liv-

ing mammals (Fig. 5:node 1). In a more derivedclade of Morganucodon, Haldanodon, and mam-malian crown group, the antero-posterior sequentialreplacement only occurred in premolars, but not inposterior molariforms (molars) (Fig. 5:node 2).Gobiconodontids have antero-posterior sequentialreplacement for anterior molariform postcanines(‘‘molars’’) but not for the posterior molars (Jenkinsand Schaff, 1988). This is consistent with the re-placement pattern of Morganucodon, Haldanodon,and most multituberculates.

The antero-posterior sequential replacement ofpremolars is a predominating pattern widespreadamong early divergent clades of the mammaliancrown group (possibly including triconodontids)and in many (but not all) placental groups (Schmid,1972; Smith, 2000). However, the trechnotherianclade (Fig. 5:node 3) is characterized by an alter-nating, diphyodont sequence (p2 → p4 → p3) asevidenced by most stem taxa of this clade (Zhan-gheotherium, Dryolestes, and Slaughteria). The bas-al eutherians retained the primitive condition of analternate, antero-posterior replacement of premolars(p2 → p4 → p3). The sequential replacement char-acteristics in the majority of extant placentals (Fig.5:node 4) is a secondarily derived condition withineutherians. Among the crown placentals, the re-placement sequence is in the antero-posterior direc-tion (p2 → p3 → p4) for most groups includingungulates and carnivores (Schmid, 1972; Smith,2000). The postero-anterior sequential replacement(p4 → p3 → p2) occurs in various insectivores, andthe majority of the prosimian primates (Smith,2000) may not represent a generalized condition forextant placentals as a whole, in contrast to a pre-vious belief. The single replacement of P3/p3 ofmetatherians (Fig. 5:node 5) is the most deriveddental replacement pattern of all therian mammals.This metatherian dental replacement is correlatedwith the high specialized reproduction seen in ex-tant marsupials.

ACKNOWLEDGMENTS

We are all indebted to our friend and colleague Malcolm C.McKenna for his inspiration for our own work on early evolutionof mammals and his friendship. We are also thankful to Drs.Mary R. Dawson and Jason A. Lillegraven for this opportunityto contribute to the Festschrift volume in honor of Malcolm Mc-Kenna. During the course of this work, we benefited from dis-cussion on various issues about mammalian dental replacementswith our friends and colleagues J. D. Archibald, K. C. Beard, J.I. Bloch, M. R. Dawson, W. P. Luckett, T. Martin, S. Olsen, G.W. Rougier, and K. K. Smith. We thank J. A. Lillegraven and

M. R. Dawson who reviewed the paper and provided very usefulcomments and editorial help. Q. Ji, S. McLaren and J. R. Wibleprovided access to comparative materials. M. A. Klingler and A.Kaim assisted with graphics. This research was supported by theInstitute of Paleobiology, PAN (Z. Kielan-Jaworowska), a col-laborative program between NSF and PAN (Z. Kielan-Jawo-rowska and R. L. Cifelli), the National Geographic Society (R.L. Cifelli and Z.-X. Luo), the National Science Foundation(NSF), USA (DEB-9870173 to R. L. Cifelli and DEB-95278902to Z.-X. Luo), and the Carnegie Museum (Z.-X. Luo).

172 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

LITERATURE CITED

ARCHER, M., P. MURRAY, S. J. HAND, AND H. GODTHELP. 1993.Reconsideration of monotreme relationships based on theskull and dentition of the Miocene Obdurodon dicksoni. Pp.75–94, in Mammal Phylogeny: Mesozoic Differentiation,Multituberculates, Monotremes, Early Therians, and Mar-supials (F. S. Szalay, M. J. Novacek and M. C. McKenna,eds.). Springer-Verlag, New York.

ARCHIBALD, J. D., AND A. O. AVERIANOV. 1997. New evidencefor the ancestral placental premolar count. Journal of Ver-tebrate Paleontology, 17(3-Supplement):29A.

BERKOWITZ, B. K. 2000. Tooth replacement patterns in non-mammalian vertebrates. Pp. 186–200, in Development,Function and Evolution of Teeth (M. F. Teaford, M. M.Smith, and M. W. J. Ferguson, eds). Cambridge UniversityPress, Cambridge.

BLOCH, J. I., D. M. BOYER, P. D. GINGERICH AND G. F. GUNNELL.2002. New primitive paromomyid from the Clarkforkian ofWyoming and dental eruption in Plesiadapiformes. Journalof Vertebrate Paleontology, 22:366–379.

BLOCH, J. I., K. D. ROSE, AND P. D. GINGERICH. 1998. New spe-cies of Batodonoides (Lipotyphla, Geolabididae) from theearly Eocene of Wyoming: smallest known mammal? Jour-nal of Mammalogy, 79:804–827.

BOWN, T. M., AND M. J. KRAUS. 1979. Origin of the tribosphenicmolar and metatherian and eutherian dental formulae. Pp.172–181, in Mesozoic Mammals: The First Two-thirds ofMammalian History (J. A. Lillegraven, Z. Kielan-Jawo-rowska and W. A. Clemens, eds.). University of CaliforniaPress, Berkeley, CA.

BRINK, A. A. 1956. Speculations on some advanced mammaliancharacteristics in higher mammal-like reptiles. Palaeontolo-gica Africana, 4:77–95.

BUTLER, P. M. 1939. The teeth of the Jurassic mammals. Pro-ceedings of the Zoological Society of London, 109:329–356.

BUTLER, P. M. 2000. Review of the early allotherian mammals.Acta Palaeontologica Polonica, 45:317–342.

BUTLER, P. M., AND B. KREBS. 1973. A pantotherian milk den-tition. Palaontologische Zeitschrift, 47:256–258.

CHURCHFIELD, S. 1990. The natural history of shrews. ComstockPublishing Associates, Ithaca, New York.

CIFELLI, R. L. 1999. Therian teeth of unusual design from themedial Cretaceous (Albian-Cenomanian) Cedar MountainFormation, Utah. Journal of Mammalian Evolution, 6:247–270.

. 2000. Counting premolars in early eutherian mammals.Acta Palaeontologica Polonica, 45:195–198.

CIFELLI, R. L., T. B. ROWE, W. P. LUCKETT, J. BANTA, R. REYES,AND R. I. HOWES. 1996. Origin of marsupial pattern of toothreplacement: fossil evidence revealed by high resolution X-ray CT. Nature, 379:715–718.

CLEMENS, W. A. 1963. Fossil mammals of the type Lance For-mation, Wyoming. Part I. Introduction and Multituberculata.University of California Publications in Geological Sciences,48:1–105.

. 1966. Fossil mammals from the type Lance FormationWyoming. Part II. Marsupialia. University of CaliforniaPublications in Geological Sciences, 62:1–122.

. 1973. Fossil mammals from the type Lance FormationWyoming. Part III. Eutheria and summary. University ofCalifornia Publications in Geological Sciences, 94:1–102.

CLEMENS, W. A., AND Z. KIELAN-JAWOROWSKA. 1979. Multitu-

berculata. Pp. 99–149, in Mesozoic Mammals: The FirstTwo-thirds of Mammalian History (J. A. Lillegraven, Z.Kielan-Jaworowska and W. A. Clemens, eds.). University ofCalifornia Press, Berkeley, CA.

CLEMENS, W. A., AND J. A. LILLEGRAVEN. 1986. New Late Cre-taceous, North American advanced therian mammals that fitneither the marsupial nor eutherian molds. Contributions toGeology, University of Wyoming, Special Paper, 3:55–85.

CROMPTON, A. W. 1963. Tooth replacement in the cynodontThrianxodon liorhinus Seeley. Annals of South Africa Mu-seum, 46:279–521.

. 1972. Postcanine occlusion in cynodonts and tritylo-donts. Bulletin of the British Museum (Natural History), Ge-ology, 21:30–71.

. 1974. The dentition and relationships of the southernAfrican Triassic mammals, Erythrotherium parringtoni andMegazostrodon rudnerae. Bulletin of the British Museum(Natural History), Geology, 24:397–437.

. 1995. Masticatory function in nonmammalian cyno-donts and early mammals. Pp. 55–75, in Functional Mor-phology in Vertebrate Paleontology (J. J. Thomason, ed.).Cambridge University Press, Cambridge.

CROMPTON, A. W., AND W. L. HYLANDER. 1986. Changes in man-dibular function following the acquisition of a dentary-squa-mosal joint. Pp. 263–282, in The Ecology and Biology ofMammal-like Reptiles (N. Hotton, III, P. D. MacLean, J. J.Roth and E. C. Roth, eds.). Smithsonian Institution Press,Washington, D. C.

CROMPTON, A. W., AND Z.-X. LUO. 1993. Relationships of theLiassic mammals Sinoconodon, Morganucodon, and Din-netherium. Pp. 30–44, in Mammal Phylogeny: MesozoicDifferentiation, Multituberculates, Monotremes, Early The-rians, and Marsupials (F. S. Szalay, M. J. Novacek and M.C. McKenna, eds.). Springer-Verlag, New York.

CROMPTON, A. W., AND P. PARKER. 1978. Evolution of the mam-malian masticatory apparatus. American Scientist, 66:192–201.

CROMPTON, A. W., AND A.-L. SUN. 1985. Cranial structure andrelationships of the Liassic mammal Sinoconodon. Zoolog-ical Journal of the Linnean Society, 85:99–119.

CUI, G.-H., AND A. SUN. 1987. Postcanine root system of trity-lodonts. Vertebrata PalAsiatica, 25:245–259.

EDMUND, A. G. 1960. Tooth replacement phenomena in the low-er vertebrates. Contributions of Royal Ontario Museum, LifeScience Division, 52:1–190.

ERICKSON, G. M. 1996. Incremental lines of von Ebner in di-nosaurs and the assessment of tooth replacement rates usinggrowth line counts. Proceedings of the National Academyof Sciences USA, 93:14623–14627.

ERICKSON, G. M., AND T. A. TUMANOVA. 2000. Growth curveand life history attributes of Psittacosaurus mongoliensis(Ceratopsia: Psittacosauridae) inferred from long bone his-tology. Zoological Journal of Linnean Society, 130:551–566.

EVANS, H. E. 1995. Miller’s Anatomy of the Dog. (3rd edition).W. B. Saunders Company, Philadelphia. i–xi, pp.1–1113.

GINGERICH, P. D., AND B. H. SMITH. 1984. Allometric scaling inthe dentition of primates and insectivores. Pp. 257–272, inSize and Scaling in Primate Biology (W. L. Jungers, ed.).Plenum Press, New York.

GOW, C. E. 1980. The dentitions of the Trithelodontidae (Ther-

2004 173LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

apsida: Cynodontia). Proceedings of the Royal Society, Lon-don. 208:461–481.

. 1985. Apomorphies of the Mammalia. South AfricanJournal of Science, 81:558–560.

. 1986. A new skull of Megazostrodon (Mammalia: Tri-conodonta) from the Elliot Formation (Lower Jurassic) ofsouthern Africa. Palaeontologia Africana, 26:13–23.

GREEN, H. L. H. H. 1937. The development and morphology ofthe teeth of Ornithorhynchus. Philosophical Transactions ofthe Royal Society of London, 288:367–420.

GREENWALD, N. S. 1988. Patterns of tooth eruption and replace-ment in multituberculate mammals. Journal of Vertebrate Pa-leontology, 8:265–277.

HAHN, G. 1978. Die Multituberculata, eine fossille Saugetier-Ordnung. Sonderband der naturwissenschaftlische Verein,Hamburg, 3:61–95.

HAHN, G., AND R. HAHN. 1998. Neue Beobachtungen an Pla-giaulacoidea (Multituberculata) des Ober-Juras 1. ZumZahn-Wechsel bei Kielanodon. Berliner Geowissenschaftlis-che Abhandlungen, E28:1–7.

HILDEBRAND, M. 1974. Analysis of Vertebrate Structure. JohnWiley & Sons, New York.

HOPSON, J. A. 1965. Tritylodontid therapsids from Yunnan andthe cranial morphology of Bienotherium. Ph.D. Dissertation,The University of Chicago.

. 1971. Postcanine replacement in the gomphodont cy-nodont Diademodon. Pp. 1–21, in Early Mammals (D. M.Kermack and K. A. Kermacke, eds.). Academic Press, Lon-don.

. 1973. Endothermy, small size, and the origin of mam-malian reproduction. American Naturalist, 107:446–452.

. 1994. Synapsid evolution and the radiation of non-eu-therian mammals. Pp. 190–219, in Major Features of Ver-tebrate Evolution (D. Prothero, ed.). The Paleontological So-ciety, Knoxville.

HOPSON, J. A., AND A. W. CROMPTON. 1969. Origin of mammals.Pp. 15–72,. in Evolutionary Biology, volume 3 (T. Dob-zhansky, M. K. Hecht, and W. C. Steere, eds.). New York:Appleton-Century-Crofts.

HU, Y.-M., Y.-Q. WANG, C.-K. LI, AND Z.-X. LUO. 1998. Mor-phology of dentition and forelimb of Zhangheotherium. Ver-tebrata PalAsiatica, 36:102–125.

HU, Y.-M., Y.-Q. WANG, Z.-X. LUO, AND C.-K. LI. 1997. A newsymmetrodont mammal from China and its implications formammalian evolution. Nature, 390:137–142.

JENKINS, F. A., JR. 1990. Monotremes and the biology of Me-sozoic mammals. Netherlands Journal of Zoology, 40:5–31.

JENKINS, F. A., JR., A. W. CROMPTON, AND W. R. DOWNS. 1983.Mesozoic mammals from Arizona: new evidence on mam-malian evolution. Science, 222:1233–1235.

JENKINS, F. A., JR., AND C. R. SCHAFF. 1988. The Early Creta-ceous mammal Gobiconodon (Mammalia, Triconodonta)from the Cloverly Formation in Montana. Journal of Verte-brate Paleontology, 6:1–24.

JI, Q., Z.-X. LUO, C-X. YUAN, J. R. WIBLE, J.-P. ZHANG, AND J. A.GEORGI. 2002. The earliest known eutherian mammal. Na-ture, 416:816–822.

KAY, R. F., AND E. L. SIMONS. 1983. Dental formulae and dentaleruption patterns in Parapithecidae (Primates, Anthropoi-dea). American Journal of Physical Anthropology, 62:363–375.

KERMACK, D. M., K. A. KERMACK, AND F. MUSSETT. 1968. The

Welsh pantothere Kuehneotherium praecursoris. Journal ofthe Linnean Society (Zoology), 47:407–423.

KERMACK, K. A. 1963. The cranial structure of the triconodonts.Philosophical Transactions of the Royal Society of London,246:83–103.

. 1967. Molar evolution in Mesozoic mammals. Journalof Dental Research, 46:792–795.

KERMACK, K. A., AND D. M. KERMACK. 1984. Evolution ofMammalian Characters. x 1 149. Croom Helm, London.

KERMACK, K. A., F. MUSSETT, AND H. W. RIGNEY. 1973. Thelower jaw of Morganucodon. Zoological Journal of the Lin-nean Society, 53:87–175.

. 1981. The skull of Morganucodon. Zoological Journalof the Linnean Society, 53:87–175.

KIELAN-JAWOROWSKA, Z. 1975. Preliminary description of twonew eutherian genera from the Late Cretaceous of Mongolia.Palaeontologia Polonica, 33:5–16.

. 1981. Evolution of the therian mammals in the LateCretaceous of Asia. Part IV. Skull structure in Kennalestesand Asioryctes. Palaeontologia Polonica, 42:25–78.

KIELAN-JAWOROWSKA, Z., AND D. DASHZEVEG. 1989. Eutherianmammals from the Early Cretaceous of Mongolia. ZoologicaScripta, 18:347–355.

KINDAHL, M. E. 1963. On the embryonic development of theteeth in the Golden Mole, Eremitalpa (Chryschloris) grantiBroom in South Africa. Archiva Zoologia, 16:97–115.

. 1967. Some comparative aspects of reduction of thepremolars in the Insectivora. Journal of Dental Research, 46:805–808.

KOBAYASHI, Y., D. A. WINKLER, AND L. J. JACOBS. 2002. Originof the tooth-replacement pattern in therian mammals: evi-dence from a 110 Myr old fossil. Proceedings of Royal So-ciety of London, 269:369–373

KRUSAT, G. 1980. Contribucao para o conhecimento da faunado Kimeridgiano da mina de lignito Guimarota (Leiria, Por-tugal). IV Parte. Haldanodon exspectatus Kuhne and Krusat1972 (Mammalia, Docodonta). Memorias dos Servicos Geo-logicos de Portugal, 27:1–79.

KUHNE, W. G. 1956. The Liassic therapsid Oligokyphus. BritishMuseum (Natural History), London.

LILLEGRAVEN, J. A. 1969. Latest Cretaceous mammals of upperpart of Edmonton Formation of Alberta, Canada, and reviewof marsupial-placental dichotomy in mammalian evolution.University of Kansas Paleontological Contributions, 50(Ver-tebrata 12):1–122.

LILLEGRAVEN, J. A., M. C. MCKENNA, AND L. KRISHTALKA. 1981.Evolutionary relationships of middle Eocene and youngerspecies of Centetodon (Mammalia, Insectivora, Geolabidi-dae) with a description of the dentition of Ankylodon (Adap-isoricidae). University of Wyoming Publications, 45:1–115.

LILLEGRAVEN, J. A., AND G. KRUSAT. 1991. Cranio-mandibularanatomy of Haldanodon exspectatus (Docodonta; Mamma-lia) from the Late Jurassic of Portugal and its implicationsto the evolution of mammalian characters. Contributions toGeology, University of Wyoming, 28:39–138.

LUCKETT, W. P. 1977. Ontogeny of amniote fetal membranes andtheir application to phylogeny. Pp. 439–516, in Major Pat-terns in Vertebrate Evolution (M. K. Hecht, P. C. Goody,and B. M. Hecht, eds.). Plenum Press, New York.

. 1993. An ontogenetic assessment of dental homologiesin therian mammals. Pp. 182–204, in Mammal Phylogeny:Mesozoic Differentiation, Multituberculates, Monotremes,

174 NO. 36BULLETIN CARNEGIE MUSEUM OF NATURAL HISTORY

Early Therians, and Marsupials (F. S. Szalay, M. J. Novacekand M. C. McKenna, eds.). Springer-Verlag, New York.

LUCKETT, W. P., AND N. HONG. 2000. Ontogenetic evidence fordental homologies and premolar replacement in fossil andextant caenolestids (Marsupialia). Journal of MammalianEvolution, 7:109–128.

LUCKETT, W. P., AND P. A. WOOLLY. 1996. Ontogeny and ho-mology of the dentition in dasyurid marsupials: developmentin Smithopsis virgininae. Journal of Mammalian Evolution,3:327–264.

LUCKETT, W. P., AND U. ZELLER. 1989. Developmental evidencefor dental homologies in the monotreme Ornithorhynchusand its systematic implications. Zeitschrift fur Saugetier-kunde, 54:193–204.

LUO, Z.-X. 1994. Sister-group relationships of mammals andtransformations of diagnostic mammalian characters. Pp.98–128, in In the Shadow of the Dinosaurs—Early MesozoicTetrapods (N. C. Fraser and H.-D. Sues, eds.). CambridgeUniversity Press, Cambridge.

LUO, Z.-X., R. L. CIFELLI, AND Z. KIELAN-JAWOROWSKA. 2001a.Dual origin of tribosphenic mammals. Nature, 409:53–57.

LUO, Z.-X., A. W. CROMPTON, AND S. G. LUCAS. 1995. Evolu-tionary origins of the mammalian promontorium and co-chlea. Journal of Vertebrate Paleontology, 15:113–121.

LUO, Z.-X., A. W. CROMPTON, AND A.-L. SUN. 2001b. A newmammal from the Early Jurassic and evolution of mamma-lian characteristics. Science, 292:1535–1540.

LUO, Z.-X., JI, Q., AND JI, S-A. 2001. Dental replacement inZhangheotherium and its implications on evolution of mam-malian dentition. Journal of Morphology, 248(3-Supple-ment):256.

LUO, Z.-X., Z. KIELAN-JAWOROWSKA, AND R. L. CIFELLI. 2002.In quest for a phylogeny of Mesozoic mammals. Acta Pa-laeontologica Polonica, 47:1–78.

MAIER, W. 1993. Cranial morphology of the therian commonancestor, as suggested by the adaptations of neonate mar-supials. Pp. 165–181, in Mammal Phylogeny: Mesozoic Dif-ferentiation, Multituberculates, Monotremes, Early Therians,and Marsupials (F. S. Szalay, M. J. Novacek and M. C. Mc-Kenna, eds.). Springer-Verlag, New York.

MARTIN, T. 1997. Tooth replacement in Late Jurassic Dryoles-tidae (Eupantotheria, Mammalia). Journal of MammalianEvolution, 4:1–18.

. 1999. Dryolestidae (Dryolestoidea, Mammalia) ausdem Oberen Jura von Portugal. Abhandlungen der Sencken-bergischen Naturforschenden Gesellschaft, 550:1–119.

MARTIN, T., AND M. NOWOTNY. 2000. The docodont Haldanodonfrom the Guimarota mine. Pp. 91–96, in Guimarota—a Ju-rassic Ecosystem (T. Martin and B. Krebs, eds.). Verlag Dr.Friedrich Pfeil, Munchen.

MCKENNA, M. C. 1975. Toward a phylogenetic classification ofthe Mammalia. Pp. 21–46, in Phylogeny of the Primates (W.P. Luckett and F. S. Szalay, eds.). Plenum Publishing Cor-poration, New York.

MCKENNA, M. C., AND S. K. BELL. 1997. Classification of Mam-mals Above the Species Level. Columbia University Press,New York.

MCKENNA, M. C., Z. KIELAN-JAWOROWSKA, AND J. MENG. 2000.Earliest eutherian mammal skull from the Late Cretaceous(Coniacian) of Uzbekistan. Acta Palaeontologica Polonica,45:1–54.

MILLS, J. R. E. 1971. The dentition of Morganucodon. Pp. 29–63, in Early Mammals (Kermack and Kermack, eds.). Zoo-

logical Journal of the Linnean Society 50, supplement 1,London.

. 1984. The molar dentition of a Welsh pantothere. Zoo-logical Journal of the Linnean Society, 82:189–205.

NESSOV, L. A., J. D. ARCHIBALD, AND Z. KIELAN-JAWOROWSKA.1998. Ungulate-like mammals from the Late Cretaceous ofUzbekistan and a phylogenetic analysis of Ungulatomorpha.Bulletin of the Carnegie Museum of Natural History, 34:40–88.

NOWAK, R. M. 1991. Walker’s mammals of the world. 5th Edi-tion. The Johns Hopkins University Press. Baltimore, Mary-land 1:1–642, 2:1–1629.

NOVACEK, M. J. 1986. The primitive eutherian dental formula.Journal of Vertebrate Paleontology, 6:191–196.

NOWOTNY, M., T. MARTIN, AND M. S. FISCHER. 2001. Dentalanatomy and tooth replacement of Haldanodon exspectatus(docodonta, Mammalia) from the Upper Jura of Portugal.Journal of Morphology, 248(3-Supplement):268.

OSBORN, J. W. 1971. The ontogeny of tooth succession in La-certa vivipera Jacquine (1787). Proceedings of Royal Soci-ety of London, B179:261–289.

. 1973. The evolution of dentitions. American Scientist,61:548–559.

. 1974a. On the control of tooth replacement in reptilesand its relationship to growth. Journal of Theoretical Biol-ogy, 46:509–527.

. 1974b. On the tooth succession in Diademodon. Evo-lution, 28:141–157.

OSBORN, J. W., AND A. W. CROMPTON. 1973. The evolution ofmammalian from reptilian dentitions. Breviora, 399:1–18.

PARRINGTON, F. R. 1936. On the tooth replacement in theriodontreptiles. Philosophical Transactions of Royal Society of Lon-don, B226:121–142.

. 1971. On the Upper Triassic mammals. PhilosophicalTransactions of the Royal Society, B261:231–272.

. 1973. The dentitions of the earliest mammals. Zoolog-ical Journal of the Linnean Society, 52:85–95.

. 1974. The problem of the origin of the monotremes.Journal of Natural History, 8:421–426.

. 1978. A further account of the Triassic mammals. Phil-osophical Transactions of the Royal Society of London,B282:177–204.

POND, C. M. 1977. The significance of lactation in the evolutionof mammals. Evolution, 31:177–199.

ROUGIER, G. W., J. R. WIBLE, AND J. A. HOPSON. 1996. Basicra-nial anatomy of Priacodon fruitaensis (Triconodontidae,Mammalia) from the Late Jurassic of Colorado, and a re-appraisal of mammaliaform interrelationships. AmericanMuseum Novitates, 3183:1–38.

ROUGIER, G. W., J. R. WIBLE, AND M. J. NOVACEK. 1998. Impli-cations of Deltatheridium specimens for early marsupial his-tory. Nature, 396:459–463.

ROMER, A. S. 1970. The Vertebrate Body (4th Edition). W. B.Saunders Company, Philadelphia.

ROWE, T. B. 1988. Definition, diagnosis, and origin of Mam-malia. Journal of Vertebrate Paleontology, 8:241–264.

. 1993. Phylogenetic systematics and the early history ofmammals. Pp. 129–145, in Mammal Phylogeny: MesozoicDifferentiation, Multituberculates, Monotremes, Early The-rians, and Marsupial (F. S. Szalay, M. J. Novacek, and M.C. McKenna, eds.). Springer-Verlag, New York.

SCHMID, E. 1972. Atlas of animal bones—for prehistorians, ar-

2004 175LUO ET AL.—EVOLUTION OF DENTAL REPLACEMENT IN MAMMALS

chaeologists and Quarternary geologists. Amsterdam, Lon-don, New York: Elsevier Publishing Company.

SIGOGNEAU-RUSSELL, D., D. DASHZEVEG, AND D. E. RUSSELL.1992. Further data on Prokennalestes (Mammalia, Eutheriainc. sed.) from the Early Cretaceous of Mongolia. ZoologicaScripta, 21:205–209.

SIMPSON, G. G. 1928. A Catalogue of the Mesozoic Mammaliain the Geological Department of the British Museum. Trust-ees of the British Museum, London.

. 1929. The dentition of Ornithorhynchus as evidence ofits affinities. American Museum Novitates, 390:1–15.

SMITH, B. H. 1994. Sequence and emergence of the permanentteeth in Macaca, Pan, Homo, and Australopithecus: Its evo-lutionary significance. American Journal of Human Biology,6:61–76.

. 2000. ’Schulz’s rule’ and the evolution of tooth emer-gence and replacement patterns in primates and ungulates.Pp. 212–228, in Development, Function and Evolution ofTeeth (M. F. Teaford, M. M. Smith, and M. W. J. Ferguson,eds). Cambridge University Press, Cambridge.

SMITH, K. K. 1996. Integration of craniofacial structures duringdevelopment in mammals. American Zoologist, 36:70–79.

. 1997. Comparative patterns of craniofacial develop-ment in eutherian and metatherian mammals. Evolution, 51:1663–1678.

SZALAY, F. S. 1965. First evidence of tooth replacement in thesubclass Allotheria (Mammalia). American Museum Novi-tates, 2226:1–12.

TYNDALE-BISCOE, H., AND M. RENFREE. 1987. Reproductivephysiology of marsupials. Cambridge University Press,Cambridge.

WANG, Y.-Q., Y.-M. HU, J. MENG, AND C-K. LI. 2001. An os-sified Meckel’s cartilage in two Cretaceous mammals andorigin of the mammalian middle ear. Science, 294:357–361.

WESTERGAARD, B., AND M. W. J. FERGUSON. 1986. Developmentof the dentition in Alligator mississippiensis. Early embry-

onic development in the lower jaw. Journal of Zoology(London), (A)210:575–597.

. 1987. Development of the dentition in Alligator mis-sissippiensis. Later development in the lower jaws of em-bryos, hatchlings and young juveniles. Journal of Zoology(London), 212:199–222.

WIBLE, J. R., G. W. ROUGIER, M. J. NOVACEK, M. C. MCKENNA,AND D. DASHZEVEG. 1995. A mammalian petrosal from theEarly Cretaceous of Mongolia: implications for the evolutionof the ear region and mammaliamorph relationships. Amer-ican Museum Novitates, 3149:1–19.

WILLIAMS, P. L., R. WILLIAMS, M. DYSON, AND L. H. BANNISTER.1989. Gray’s Anatomy (37th Edition). New York: ChurchillLivingstone.

WOODBURNE, M. O., AND R. H. TEDFORD. 1975. The first Tertiarymonotreme from Australia. American Museum Novitates,2588:1–11.

YANG, Z-J. (YOUNG, C. C.) 1982. Selected Works of YangZhongjian (Young Chung-Chien). (Edited by Editorial Com-mittee for Selected Works of Yang Zhongjian, Chaired byM. C. Chow [M.-C. Zhou]) Science Press, Beijing.

ZELLER, U. 1999. Mammalian reproduction: origin and evolu-tionary transformations. Zoologischer Anzeiger, 238:117–130.

ZELLER, U., AND C. FREYER. 2001. Early ontogeny and placen-tation of the grey short-tailed opossum, Monodelphis do-mestica (Didelphidae: Marsupialia): contribution to the re-construction of the marsupial morphotype. Journal of Zoo-logical Systematics and Evolutionary Research, 39:137–158.

ZHANG, F.-K., A. W. CROMPTON, Z.-X. LUO, AND C. R. SCHAFF.1998. Pattern of dental replacement of Sinoconodon and itsimplications for evolution of mammals. Vertebrata Pal-Asiatica, 36:197–217.

ZIEGLER, A. C. 1971. A theory of the evolution of therian dentalformulas and replacement pattern. Quarterly Review of Bi-ology, 46:226–249.